On the basis of the assumptions of statistical theories, nonlinear expressions for some of the fundamental equations of nucleation theory have been derived. The results apply to any of the statistical theories but are specifically applied to the modification of the original Lothe–Pound theory. In the range of usual supersaturations the numerical results of the linearized Lothe–Pound theory (LLP) differ only slightly from those of the nonlinear Lothe–Pound theory (ELP), but the difference increases with degree of supersaturation. All fundamental equations of nucleation theory can be expressed in terms of a dimensionless energy parameter that also defines the condition of onset of spontaneous condensation without nucleation, expected to occur at higher supersaturations. Furthermore, under the premises of the present theory the growing clusters will pass an intermediate stable stage at a specific radius before reaching the size of a stable critical nucleus. This effect should result in changes in the relaxation time for establishing steady-state of nucleation.

A molecular dynamics computer simulation technique of model systems was used to (a) determine the extent of coalescence of a dispersion of adatoms on the amorphous surface as a function of the adatom-substrate interaction energy, and (b) evaluate the effect of the adatom-substrate interaction energy on the mobility of a 10 atom cluster on amorphous and crystalline surfaces. Results show that at low temperatures the diffusion of adatoms on the amorphous surface is limited by the high density of sites from which adatoms require a significant amount of thermal energy in order to escape. At higher temperatures, the amorphous surface can relax and accommodate adatoms and clusters more readily than a low index crystalline surface, again inhibiting diffusion. Such results help explain the experimentally observed high sticking coefficients and higher densities of small clusters experimentally observed in depositions on amorphous surfaces as compared to crystalline surfaces.

A molecular dynamics study of surface self-diffusion on four different crystal surfaces was accomplished for atoms interacting via a Lennard-Jones 12-6 potential energy function. Both the surface and adsorbed atoms experienced realistic thermal motion which included all anharmonic effects. The formation of vacancy-ad-atom defect pairs leading to surface layer melting was observed below the bulk melting temperature for the (111), (113), (110) and (100) surfaces. Diffusion activation energies and exponential prefactors were calculated. Details for unusual atomic transport mechanisms on the (111) and especially (110) surfaces are also reported. Additional features include the analysis of velocity correlations and vibrational spectra associated with different types of atomic motion at surfaces.

In the diffusional creep of a substitutional solid solution where the diffusivities of the components are different, there is a transient state in which the vacancy flux uses preferentially those atoms with the higher diffusivity, producing segregation of the components. A mathematical formulation of steady-state diffusional creep of a multicomponent solid solution is obtained and, by using a onedimensional model of diffusional creep, the amount of segregation required to achieve the steady state in a binary solution is evaluated in terms of either the activity coefficients or the thermodynamic factor. A simple rule of thumb to evaluate the amount of the segregation is given and results for ideal and regular solid solutions are obtained.

Molecular dynamics was used to study the structure, dispersion and short-time behavior of ten-atom clusters adsorbed onto amorphous and crystalline substrates, in which the cluster atoms differed from the substrate atoms. Two adatom–substrate model systems were chosen; one, in which the interaction energy between adatom pairs was greater than that between substrate pairs, and the other, in which the reverse was true. At relatively low temperature ranges, increased dispersion of cluster atoms occurred: (a) on the amorphous substrate as compared to the FCC(100) surface, (b) with increasing reduced temperature, and (c) with adatom–substrate interaction energy stronger than adatom–adatom interaction. Two dimensional clusters (rafts) on the FCC(100) surface displayed migration of edge atoms only, indicating a mechanism for the cluster rotation and shape changes found in experimental studies.

A monatomic amorphous surface has been simulated above and below the gradual melting transition using molecular dynamics for atoms interacting through a Lennard-Jones 12−6 potential energy function. Detailed atomic trajectories are presented and averaged to yield diffusion constants and activation energies. Surface melting, which occurs at lower temperatures than for the amorphous bulk and most crystalline surfaces, is described in terms of radial distribution functions, velocity correlations, and vibrational spectra.

The validity of the relation ..delta..S = -(par. deltaT/par. delta..delta..G)/sub t/ for the entropy changes in solids was studied. It was found that the structure must be a point or state function of the variable Temperature T/sub 1/ and the tractions, t. (FS)

The equilibrium vacancy and solute distribution around an edge dislocation in a non-ideal, substitutional, multicomponent solid solution of arbitrary composition is determined. The equilibrium condition that the free energy of the system should remain unchanged when an atom of any component located inside the elastic field of the dislocation exchanges positions with a vacancy located far from the dislocation is used. A set of equilibrium conditions whose simultaneous solution gives the equilibrium vacancy and solute distribution is found. The well-known formulae for ideal and dilute binary solutions are recovered as special cases of the general formulation. The equations governing the steady-state vacancy and solute distribution around a gliding edge dislocation are found for a non-ideal, non-dilute substitutional binary solid solution. By considering the fact that no sources or sinks for vacancies are present in the field of a gliding edge dislocation, the equation governing the flux of solute atoms in the frame of reference of the moving dislocation is found. The relevant diffusion coefficient for solute-drag-controlled creep is obtained and the boundary value problem to be solved to find the steady-state solute distribution around the dislocation is stated. The solute drag force on the gliding dislocation is obtained from the power dissipated by the diffusional process.

In this paper it is shown that the equation for the entropy of activation for dislocation motion, derived previously by the authors, is in accordance with the classical definition of that quantity. The proof is presented in three stages. First, the entropy to form a dislocation in a linear elastic solid, to which no external tractions are applied, is calculated using both approaches. It is shown that both methods yield the same result. In the second stage, the entropy associated with the motion of a dislocation from a ground to an activated state, in a solid subjected to a uniform applied stress, is formulated. It is proven that the entropy of activation in this case is given by the difference between the formation entropies of the ground and activated states. This result illustrates that the entropy of activation is determined entirely by the internal stress states, and is independent of stresses produced by the applied tractions. Finally, a more general proof of the identity of the two methods is presented.

A model of the surface of Pd80Si20 metallic glass is presented. The bulk simulation is similar to Gaskell's procedure. The roughness of the surface is described in terms of coordination numbers (CN), the distribution of CN ranging from 4 to 11, with the average value 5.96. This distribution is compared to the distribution function for the surface atoms in small particles and high-Miller indexed planes. The amorphous surface is displayed via computer graphics. The roughness of the surface is also described by a fractal dimension, D, which was found to be 2.3 (the value for the Brownian island). A theoretical reconstruction of the surface was carried out under the assumption that on the clean surface the bond lengths between the atoms in the first and second layer are shorter than the bond lengths in the bulk. The reconstruction was considered as a function of coordination number. It was found that the lower the coordination of a particular site, the greater the tendency for surface segregation. In the case of Pd80Si20 the segregation of silicon takes place preferentially at sites with coordination numbers 4 to 6. The vacuum model has been extended by considering physical adsorption of helium on the glass surface. The helium adsorption did not introduce any reconstruction.

The use of a newly combined ultrahigh vacuum technique for studying continuous and particulate evaporated thin films using thermal desorption spectroscopy (TDS)-transmission electron microscopy (TEM)-transmission electron diffraction (TED) is discussed. It is shown that (1) CO thermal desorption energies of epitaxially deposited (111) Ni and (111) Pd surfaces agree perfectly with previously published data on bulk (111) single crystals, (2) contamination and surface structural differences can be detected using TDS as a surface probe and TEM as a complementary technique and (3) CO desorption signals from deposited metal coverages of one-thousandth of a monolayer should be detectable. These results indicate that the chemisorption properties of supported “microsurfaces” of metals can now be investigated with very high sensitivity. The combined use of TDS and TEM-TED experimental methods is a very powerful technique for fundamental studies in basic thin film physics and in catalysis.

A model for predicting the surface segregation of solute in very dilute binary solid alloys is developed. The two main factors contributing to the driving force for segregation are the bond strength ratio ϵ∗ and size ratio σ∗ for the solute atom in the solvent matrix. These differences are accounted for implicity by considering various solid solution systems in bulk and surface configurations and by minimizing the total potential energy of the systems (0 K) through atomic relaxation consistent with the assumed long-range, pairwise interactions between the atoms. In previous studies, these two factors have been treated separately in an “ad hoc” manner and the strain energy due to the odd-size solute atom in the solvent lattice has been estimated using various continuum elasticity models. We demonstrate the short-comings of the conventional approach, in particular the failure of the simple continuum elasticity theory to accurately estimate the elastic driving force. We invent the ϵ∗−σ∗ representation which greatly facilitates a comparison of the various theories with one another and with experiments, the usefulness of this representation residing in the important finding that the theoretical boundary separating segregation and non-segregation regions depends only on ϵ∗ and σ∗. Comparison of our theory with experiment in 31 cases yields 28 correct predictions.

The activation energies for diffusion were determined for gold, platinum and iridium adatoms on (110) and (311) Pt surfaces and were found to be in good agreement with the measurements reported by Bassett and Webber. The Lennard-Jones pair potentials were used to model the interatomic forces, and relaxation of the substrate atoms in near proximity to the adatom was considered in detail. The present calculations clarify the mechanism of the observed two-dimensional diffusion of platinum and iridium atoms on a (110) Pt surface. The results are compared with those obtained using Morse potential functions and different relaxation techniques.

Using structure factor measurements from molecular dynamics experiments on fluid phase stability, we establish the existence of the spinodal curve.(AIP)